Is throughput in constant terms across the whole window? (edited to fix a unit)
Confirmed on my side too.
Duplicated.
The units are missing from the legend.
Worth comparing to the parallel run.
That's the way to do it. (edited to fix a unit)
Nice — the Q4 view helps.
Can we see the Nordics on the same scale?
Confirmed on my side too.
Moving this to the Q1 agenda.
Good catch.
What happened to India around 2005?
Matches our numbers too.
For background: Norway changed reporting in 2006, which is why H2 looks odd.
raised this on the other chart too, same a currency effect.
Good catch.
Good call.
Thanks for pulling this together.
Spot on.
Nice — the H1 view helps.
Fair enough.
Moana Chen Should we stratify by device type? See /u/plg_moana_chen/p/plot-0005.
Moana Chen Can we add confidence bands?
The 2010 tick is off.
headcount here is inverted; it should be seasonally adjusted.
Looks good.
I'm fully on board with the decision to use the external reference as the ground truth — it gives us an objective baseline when there's ambiguity in our internal methods. See /u/plg_moana_chen/p/plot-0005.
Minor: throughput is double-counted in the tooltip.
Does this account for time zone differences? See /u/plg_moana_chen/p/plot-0005.
That matches what I had. (edited to fix a unit)
Are these deduplicated?
The 2005 tick is mislabelled.
Confirmed. See /u/plg_moana_chen/p/plot-0005.
Same conclusion here.
Confirmed.
Moana Chen Sounds right.
Works for me.
The tooltip is cut off. See /u/plg_moana_chen/p/plot-0005.
Does this include test users?
Thanks for catching that détail.
Can we label the legend? Hard to read otherwise.
Yuki Choi Good catch. (edited to fix a unit)
Moana Chen The conversion includes failed attempts.
+1, and India looks the same way.
Should we use a stacked view?
Careful,
value changed definition in 2019.Needs a legend.
Moana Chen The resolution is too coarse.
Nice — the Q1 view helps.
Works for me.
Part of the monthly rebalancing.
No.
Moana Chen Are these cumulative or rolling? See /u/plg_moana_chen/p/plot-0005.
What is the refresh cadence on this?
Makes sense, thanks.
What happened to China around 2018?
Noted.
the left axis looks stale.
I had this wrong earlier. APAC is fine; it was
series_id that was off.That's the way to do it.
Context for anyone new: retention is only comparable seasonally adjusted.
Confirmed on my side too.
I really appreciate how you've documented the edge cases here — the distinction between deleted records and deactivated ones matters more than most people realize in retention calculations.
Excellent presentation.
Which vintage?
Right — population was the part I missed.
do you know whether 2010 was restated?
Title says Q4 but the data runs longer.
Right — unemployment was the part I missed.
Agreed.
Units?
Can we label the tick labels? Hard to read otherwise.
This lines up with the other chart if you put headcount per capita.
The bars are hard to distinguish.
Decimals are inconsistent between Norway and Japan.
Is this seasonality or a structural change?
Units?
Needs a legend.
Boris Iqbal Worth checking: are there regional différences in how the metric is calculated, or are we applying a unified définition across all zones and endpoints? See /u/plg_moana_chen/p/plot-0005.
Not quite — the 2007 figure is a definition change.
Should we use a darker background? See /u/plg_moana_chen/p/plot-0005.
Very good.
Should internet penetration be as a share of total for this comparison to mean anything?
Roger that.
Nice — the H2 view helps.
Same conclusion here.
Fully aligned.
Can we get automated alerts set up?
The way you've structured the lookback window to exclude the migration period is clever — it gives us clean apples-to-apples comparisons without having to build a manual adjustment factor.
Context for anyone new: throughput is only comparable per capita.